What Microsoft Fabric Is and Why It Matters
Microsoft Fabric, launched in general availability in November 2023, is a unified analytics platform that consolidates data engineering, data integration (formerly Azure Data Factory), data warehousing (Synapse), real-time analytics, and Power BI under a single product and billing model. The commercial significance for enterprise buyers is that Fabric represents Microsoft's attempt to consolidate multiple separate Azure data services — each with independent pricing — into a single platform with a capacity-based cost model.
For organisations already using Azure Synapse, Azure Data Factory, and Power BI Premium, Fabric is both a product evolution and a potential commercial restructuring. Microsoft is actively migrating Power BI Premium capacity (P SKUs) to Fabric F SKUs — P SKU availability for new purchases ended in 2024, and existing P SKU customers are being transitioned to Fabric equivalents. This makes understanding Fabric licensing urgent for any organisation with a significant data and analytics estate on Microsoft technology.
Fabric's capacity model creates a shared resource pool that can appear economical at small scale but requires careful capacity planning at enterprise scale — where uncontrolled workload growth can exhaust capacity and trigger expensive auto-scaling behaviour.
The F SKU Capacity Licensing Model
Fabric is licensed through F SKUs — capacity units that represent compute and memory resources shared across all Fabric workloads running in a given capacity. The F number indicates Capacity Units (CUs): F2 provides 2 CUs; F64 provides 64 CUs; F2048 provides 2,048 CUs. Each CU represents a baseline of processing and memory resource that all Fabric experiences — Lakehouse, Warehouse, Pipelines, Power BI datasets, Real-Time Analytics, and Notebooks — draw from concurrently.
| F SKU | Capacity Units | Approx. Monthly Cost (List) | Power BI Premium Equivalent |
|---|---|---|---|
| F2 | 2 CUs | ~$262/month | None (sub-P1) |
| F4 | 4 CUs | ~$524/month | None (sub-P1) |
| F8 | 8 CUs | ~$1,049/month | None (sub-P1) |
| F16 | 16 CUs | ~$2,098/month | None (sub-P1) |
| F32 | 32 CUs | ~$4,196/month | None (sub-P1) |
| F64 | 64 CUs | ~$8,393/month | ~P1 equivalent |
| F128 | 128 CUs | ~$16,786/month | ~P2 equivalent |
| F256 | 256 CUs | ~$33,573/month | ~P3 equivalent |
| F512 | 512 CUs | ~$67,146/month | ~P4 equivalent |
These list prices represent pay-as-you-go or monthly reservation rates. Reserved capacity commitments (annual or multi-year) reduce costs by approximately 40% versus pay-as-you-go. Enterprise Agreement customers with MACC commitments can apply Fabric consumption to their MACC balance and negotiate additional discounts as part of the overall EA package.
The critical operational consideration with F SKU capacity is workload isolation. A single F64 capacity shared across a large analytics team with concurrent heavy Lakehouse processing, Power BI report rendering, and pipeline execution will exhaust CU allocation faster than a dedicated Power BI Premium P1 used solely for report delivery. Capacity planning for Fabric requires modelling concurrent workload demand across all Fabric experiences, not just the Power BI reporting load that organisations are familiar with sizing from the Premium P SKU era.
Fabric vs Power BI Premium: Migration Path
For organisations on Power BI Premium P SKUs, the migration to Fabric F SKUs is effectively mandatory — Microsoft ended new P SKU sales in 2024 and has communicated a sunset roadmap for existing P SKU customers. The commercial transition is broadly neutral: F64 provides comparable Power BI Premium functionality to P1, at similar pricing (F64 at approximately $8,393/month versus P1 at approximately $4,995/month at list, before EA discounts). The price delta reflects Fabric's expanded workload coverage — the F SKU provides access to all Fabric experiences, not just Power BI Premium.
For organisations that exclusively use Power BI Premium for reporting and dashboarding without planning to expand to Lakehouse, data engineering, or Synapse Analytics, the migration from P1 to F64 represents a price increase for equivalent functionality. This migration moment is a commercial event that should be managed within the context of your broader Microsoft EA renewal — using the migration as leverage to extract additional concessions on M365 pricing or Azure commitments in exchange for committing to Fabric at scale.
OneLake Storage: The Hidden Cost Layer
OneLake is the unified storage layer underlying all Fabric workloads. All data ingested into Fabric — whether through Pipelines, Lakehouse, Warehouse, or direct upload — is stored in OneLake, which is built on Azure Data Lake Storage Gen2 and billed separately from Fabric capacity.
OneLake storage is priced at approximately $0.023/GB/month for standard storage (list price, before EA or MACC discounts). For data-intensive organisations — financial services, healthcare, manufacturing, retail — with petabyte-scale analytics estates, OneLake storage becomes a material monthly cost that can exceed the Fabric capacity cost itself at large data volumes.
The key OneLake cost management disciplines are: implementing data retention policies that archive or delete data beyond its analytical useful life; tiering less-accessed data to Azure Data Lake Storage cool or archive tiers rather than OneLake hot storage; and preventing unnecessary data duplication between OneLake and existing Azure storage infrastructure that predates Fabric adoption. Organisations that migrate their existing Azure Data Lake Gen2 storage to OneLake (which is architecturally straightforward) can consolidate billing without duplicating storage cost — but those that run parallel estates effectively pay twice.
User Licensing Within Fabric Capacity
Within a Fabric capacity, users interacting with Fabric content — viewing Power BI reports, accessing Lakehouse data, running notebooks — require licensing. The model varies by access type:
Power BI Pro users (approximately $10/user/month EA) can access Power BI content hosted in Fabric capacities and interact with Fabric data in permitted ways. For organisations with significant existing Power BI Pro deployments, Fabric capacity extends the value of those licences rather than replacing them.
Premium Per User (PPU) licences (approximately $20/user/month) provide Power BI Premium features on a per-user basis — as an alternative to capacity-based F SKUs for smaller deployments or specialised user populations. For organisations with fewer than 50–100 active premium users, PPU is often more economical than F64 capacity; above that threshold, the per-user cost of PPU typically exceeds the cost of shared F SKU capacity.
Users with M365 E5 or F3 licences including Power BI may already have base Power BI capabilities included — but Fabric-specific workloads (Lakehouse, Data Engineering, Synapse) require Fabric capacity regardless of M365 licence tier.
Incorporating Fabric into Your EA Negotiation
Microsoft Fabric creates a significant EA negotiation opportunity for organisations making the transition from legacy Azure data services or Power BI Premium. The migration moment — particularly the P SKU to F SKU transition — is a commercial event that Microsoft's account team will want to manage proactively. Used strategically, this creates negotiating leverage you should exploit.
The core negotiation tactics for Fabric within an EA context are: bundling Fabric capacity commitments with broader EA or MACC renewals to negotiate unified discounting across the Microsoft data platform; using Databricks, Snowflake, or Google BigQuery as competitive alternatives to anchor the negotiation (Fabric displaces these platforms directly, and Microsoft will price competitively when a credible alternative evaluation is under way); negotiating annual escalator caps on Fabric capacity reservations, which Microsoft has shown willingness to include for enterprise-scale commitments; and ensuring any Fabric commitment includes provisions for downward capacity adjustment should actual utilisation not match the modelled requirement.
Organisations that negotiate Fabric as part of a broader Microsoft relationship review — rather than as a standalone procurement — consistently achieve 20–35% better pricing than organisations that accept the standard reservation pricing.
Total Cost of Ownership Considerations
The TCO of Microsoft Fabric requires modelling four distinct cost components: Fabric capacity (F SKU monthly cost), OneLake storage (volume × $0.023/GB/month), user licences (Power BI Pro for all content consumers), and Azure networking costs (egress from OneLake to on-premises or third-party systems). At enterprise scale, all four components are material and all four carry negotiation potential within an EA or MACC framework.
The displacement value of Fabric — replacing Azure Synapse Analytics, Azure Data Factory, and Azure Data Lake Storage with unified Fabric billing — is frequently used by Microsoft's account teams to justify Fabric adoption on TCO grounds. We recommend organisations independently verify this displacement value with actual cost modelling rather than accepting Microsoft's account team narrative. In our experience, the displacement savings are real for organisations that fully retire the replaced services — but partial adoption of Fabric alongside continued use of Synapse, Data Factory, and ADLS creates cost addition rather than displacement, increasing overall Microsoft data platform spend.
For a complete framework on negotiating your Microsoft relationship, see our Microsoft EA Negotiation Guide 2026 and download our Microsoft EA Guide white paper.